Proc. Nati. Acad. Sci. USAVol. 84, pp. 7681-7685, November 1987Medical Sciences

Engraftment of a clonal bone marrow stromal cell line in vivostimulates hematopoietic recovery from total body irradiationPERVIN ANKLESARIA*, KENNETH KASE*, JULIE GLOWACKIt, CHRISTIE A. HOLLAND*,MARY ANN SAKAKEENY*, JOCYNDRA A. WRIGHT*, T. J. FITZGERALD*,CHI-YU LEEO, AND JOEL S. GREENBERGER*§*Department of Radiation Oncology, University of Massachusetts Medical School, Worcester, MA 01605; tDepartment of Surgery, Childrens Hospital,Boston, MA 02115; and tDepartment of Gynecology, University of British Columbia, Vancouver, BC, Canada V6T 2B5

Communicated by E. Donnall Thomas, June 11, 1987

ABSTRACT Whether bone marrow stromal cells of do-nors contribute physiologically to hematopoietic stem cellreconstitution after marrow transplantation is unknown. Todetermine the transplantability of nonhematopoietic marrowstromal cells, stable clonal stromal cell line (GB1/6) expressingthe a isoenzyme of glucose-6-phosphate isomerase (Glu6PI-a,D-glucose-6-phosphate ketol-isomerase; EC 5.3.1.9) was de-rived from murine long-term bone marrow cultures and maderesistant to neomycin analogue G418 by retroviral gene trans-fer. GB1/6 cells were fibronectin+, laminin', and collagen-type n+ and collagen type 1-; these GB1/6 cells supported invitro growth of hematopoietic stem cells forming colony-forming units of spleen cells (CFU-S) and of granulocytes,erythrocytes, and macrophage/megakarocytes (CFU-GEMM)in the absence of detectable growth factors interleukin 3(multi-colony-stimulating factor), granulocyte/macrophagecolony-stimulating factor, granulocyte-stimulating factor, ortheir poly(A)+ mRNAs. The GB1/6 cells produced macrophagecolony-stimulating factor constitutively. Recipient C57BL/6J(glucose-6-phosphate isomerase b) mice that received 3-Gytotal-body irradiation and 13 Gy to the right hind limb wereinjected i.v. with GB1/6 cells. Engrafted mice demonstrateddonor-originating Glu6PI-a+ stromal cells in marrow sinuses insitu 2 mo after transplantation and a significantly enhancedhematopoietic recovery compared with control irradiatednontransplanted mice. Continuous (over numerous passages)marrow cultures derived from transplanted mice demonstratedG418-resistant, Glu6PI-a+ stromal colony-forming cells andgreater cumulative production of multipotential stem cells ofrecipient origin compared with cultures established fromirradiated, nontransplanted control mice. These data areevidence for physiological function in vivo of a transplantedbone marrow stromal cell line.

Successful transplantation of bone marrow stem cells relieson their ability to proliferate and differentiate in contact withstromal cells of the microenvironment. Total-body irradia-tion (TBI) and/or treatment with chemotherapeutic alkylat-ing agents before transplantation of autologous or allogeneicstem cells does not always allow hematopoietic recovery dueto functional change in the recipient's marrow stromal cells(1). Correcting the defective marrow stroma by transplanta-tion of marrow stromal cells is one theoretical therapy.Several earlier reports suggest that marrow stromal cells canbe transplanted in vivo. Werts et al. (2) reported thatirradiated mouse limbs were reconstituted by stromal cellprogenitors migrating through the circulation. Donor-origi-nating marrow fibroblasts have been detected in lethallyirradiated mice after i.v. bone marrow transplantation (2-4).Long-term marrow cultures (LBMCs) established from hu-

man marrow transplant patients, revealed donor-originatingfibroblastic and endothelial cells (5). In contrast, otherstudies have failed to show donor-originating stromal cellsafter bone marrow transplantation (6-8).

MATERIALS AND METHODSDerivation and Characterization of Stable Marrow Stromal

Cell Lines and Test of Physiological Function in Vitro. Toestablish a stromal cell line with a dominant selectablemarker, the neomycin-resistance gene was transferred by aretroviral vector (10) to a cell line GB1/6 established from theadherent layer ofLBMCs from B6Cast (Glu6PI-a) mice (9) asdescribed (11). A subclone GBlneor containing the neor genewas selected in G418 (500 jig/ml) and expanded in vitro.Staining for alkaline phosphatase, acid phosphatase, perox-idase, a-naphthyl esterase, and lysozyme was done as de-scribed (12). Antisera to extracellular matrix proteins lami-nin, fibronectin (Collaborative Research, Waltham, MA),collagen types I and IV, and the alloenzyme marker Glu6PI-a(13) identified each protein in the GB1/6 cell line using eachspecific antiserum and the immunoperoxidase technique (14).For immunohistological studies, proximal tibiae were col-lected, split longitudinally, and fixed for 18 hr in 2%paraformaldehyde/0.1 M cacodylate buffer, pH 7.4. Boneswere decalcified in 0.3 M EDTA/0.1 M cacodylate buffer, pH7.4, for 4 days. Paraffin-embedded bones were sectioned at5 ,um and stained for reactivity with rabbit antiserum againstmouse Glu6PI-a and indirect immunoperoxidase staining(14). Total cellular RNA isolated from the GBlneor stromalcell line was used to prepare poly(A)+ mRNA by a modifi-cation ofthe guanidine hydrochloride extraction method (15).Specific message for interleukin 3 (IL-3) (16), granulocyte/macrophage colony-stimulating factor (GM-CSF) (17), mac-rophage colony-stimulating factor (M-CSF) (18), IL-1 (19),and granulocyte colony-stimulating factor (G-CSF) (20) wasidentified by hybridization with specific cDNA probes (>108cpm/,4g) as described (15). Methods for marrow transplan-tation in vitro and hematopoietic cell assays have also beendescribed (11, 21-22).

Total Body and Hind Limb Irradiation, Transplantation,and Measurement of Hematopoietic Recovery in Vivo. Adultrecipient C57B1/6 (Glu6PI-b) mice received TBI (3.0-8.5Gy) and in addition right-hind-limb (RHL) irradiation of10.0-12.5 Gy, delivered by a 137Cs y cell 40 irradiator. Doserate for the TBI ranged from 0.25 to 1.0 Gy/min, while dose

Abbreviations: Glu6PI-a, glucose-6-phosphate isomerase a; IL,interleukin; LBMCs, long-term bone marrow cultures; CFU, colony-forming unit; CFU-S, CFU of spleen cells; CFU-GEMM, CFU ofgranulocytes, erythrocytes, and macrophage/megakaryocytes;CSF, colony-stimulating factor; GM-CSF, granulocyte/macrophageCSF; M-CSF, macrophage CSF; G-CSF, granulocyte CSF; TBI,total-body irradiation; RHL, right hind limb.§To whom reprint requests should be addressed.

7681

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 5,

202

1

7682 Medical Sciences: Anklesaria et al.

rate to the exposed RHL ranged from 1.0 to 1.15 Gy/min. Asingle-cell suspension of stromal cell line GB1/6 or GBlneorwas injected i.v. 48 hr after irradiation. Peripheral blood frommice transplanted with GB1/6 cell line and from controlirradiated nontransplanted mice was analyzed weekly for atleast three mice per group. No mice were bled more than onceevery 3 weeks. Red and white blood cells, hematocrit, andplatelets were counted using an automated TOA-II Sysmexcounter (American Scientific Products, Stone Mountain,GA).

RESULTS

Characterization of the GB1/6 Stromal Cell Line. Analysisof confluent cultures by phase microscopy often showedlarge binucleate cells. The presence of the Glu6PI-a isoen-zyme was identified in 100% of the GB1/6 cells using specificantiserum. GB1/6 stromal cells were also assayed for expres-sion of cytoplasmic enzymes and were positive for a-

naphthyl acetate esterase and acid phosphatase and negativefor peroxidase and alkaline phosphatase (data not shown).Extracellular matrix components produced by GB1/6 cellsincluded fibronectin, laminin, and collagen type IV, with nodetectable collagen type I (data not shown). When these datafor cell line GB1/6 were compared with several other stromalcell lines including MBA-1, MBA-13, 14F-1 (23), and D2XRII(10), the cell line was classified as endothelial-like based onpublished criteria (23). GBineor cells had cytochemicalproperties and extracellular matrix proteins indistinguishablefrom those of GB1/6 cells. Newborn C57BL/6J and subleth-ally irradiated adult C57BL/6J mice injected with 5 x 106 to1 x 107 GB1/6 cells did not show detectable tumors after 6mo (data not shown).GB1/6 cells were tested for support of purified hemato-

poietic progenitor cells in vitro. Fig. 1 shows that hemato-poietic progenitor cells were supported over 4 weeks. Thecumulative number of CFU-S-forming progenitors producedwas 23.1 + 6.1 per flask and colony-forming unit of granu-locyte, erythrocyte, macrophage/megakaryocytes (CFU-GEMM)-forming progenitors was 81.4 ± 36.2 x 102 per flask(Fig. 1). Further, the GB1/6 cell line supported growth ofIL-3-dependent multipotential hematopoietic progenitor cellline B6SUtA (17) without added IL-3 (data not shown).

0s

cz20 80 '

; , ,104

40

7 14 21 30 7 14 21 30 >Days after engraftment U

FIG. 1. In vitro support and maintenance of hematopoieticprogenitor cells by the GB1/6 cell line. To confluent plateau-phaseGB1/6 stromal cell line 2 x 107 purified day-40 hematopoietic cellsharvested from donor C57BL/6J LBMCs were engrafted. Controldonor hematopoietic progenitor cells plated without GB1/6 cell linedid not form any stromal cell colonies and were not viable at day 7(o). Weekly nonadherent cells produced from engrafted cultures (-)were harvested and assayed for the following: (i) CFU-S-formingprogenitor cells (27). CFU-S colonies were scored on day 14. Resultsare expressed as cumulative mean + SEM of colonies per flask from10 mice per week. Control mice had less than 0.1 ± 0.03 CFU-S. (ii)Multilineage progenitor cells forming CFU-GEMM colonies in re-

sponse to 10% medium conditioned by pokeweed mitogen-stimulatedspleen cells and 2.5 units of erythropoietin per ml. Results areexpressed as cumulative mean ± SEM of colonies per flask. Thedonor cell inoculum initially contained 166 + 70 CFU-S-formingprogenitors and 49.5 ± 0.5 x 102 CFU-GEMM progenitors per flask.

To determine whether the hematopoietic support capacityof GB1/6 and GBlneor cells was due to production of aknown colony-stimulating factor (CSF), and in particularIL-3, poly(A)+ mRNA from GBlneor cells was analyzed byRNA blot hybridization, for mRNA ofknown growth factors.GBlneor cells had no detectable poly(A)+ mRNA for IL-3,GM-CSF, G-CSF, or IL-1; but the cells had detectablepoly(A)+ mRNA for M-CSF (data not shown). Thus, thehematopoietic support capacity of GB1/6 cell line in vitrocould not be attributed to the synthesis of detectable quan-tities of any known hematopoietin with multi-CSF activity.Homing and Function of GB1/6 Cells in Transplanted Mice.

The ability of injected stromal cells to "home" and stablyseed into marrow sinuses in vivo was first evaluated by in vivoimmunohistochemical technique. Glu6PI-a' stromal cellswere identified in situ 2 mo after transplant in the RHLmarrow sinuses of transplanted mice. Two months aftertransplantation neither engrafted nor irradiated nonengraftedcontrol mice demonstrated detectable donor-originatingGlu6PI-a cells in spleen, liver, lung, or peritoneal washings(data not shown). As shown in Table 1 and Fig. 2 Upper, 1mo after transplantation 26.2 ± 3.2% of the adherent stromalcells in marrow cultures explanted from transplanted micewere of donor origin. The highest percentage of donor-originating cells was 82.5 ± 0.5% and 62.5 ± 12.5% of totaladherent cells in marrow explants 2 and 3 mo, respectively,after transplantation (Table 1). Glu6PI-a+ cells composed78.0% of adherent cells in LBMCs established from RHLs(13-Gy irradiated) of transplanted mice (Table 1). In contrast,nonadherent hematopoietic progenitor cells harvested fromthese same LBMCs had no detectable Glu6PI-a+ cells. Nodetectable Glu6PI-a+ stromal cells were identified in situ orin LBMCs from control-irradiated nontransplanted mice(Fig. 2 Lower, Table 1).

Explanted marrow cells from control-irradiated nontrans-planted mice and mice transplanted with GBlneor cells wereselected for growth in G418 (500 ,ug/ml). Table 2 shows thatG418-resistant stromal cell colonies were found in theexplanted RHL marrow of transplanted, but not of control-irradiated mice.The physiological function in vitro of transplanted GB1/6

stromal cells was next evaluated. At monthly intervalsLBMCs were established individually from each hind limb ofGB1/6-transplanted and -irradiated nonengrafted control

Table 1. Identification of donor-originating stromal cells in bonemarrow explanted from transplanted mice

Donor-originating Glu6PI-a+ cells, %

GB1/6 Adherent stromal cell AdherenttTime, cells injected explants* per limb, % from rightmo per mouse Right Left limb LBMCs

1 0 <0.01 <0.01 Not tested1 x 106 26.2 + 3.2 28.4 + 4.4 Not tested

2 0 <0.01 <0.01 <0.011 x 105 75.5 ± 9.5 65.0 ± 2 76.7 ± 15.75 x 105 82.5 ± 0.5 32.5 ± 3.5 78.0 ± 01 x 106 59.0 ± 14.4 16 ± 0 Not tested

3 0 <0.01 <0.01 <0.011 x 106 62.5 ± 12.5 \<0.01 Not tested

*Adherent stromal cell explants were established, and donor-originating stromal cells were identified at day 18 using specificantiserum against Glu6PI-a alloenzyme marker (13), and immuno-peroxidase staining (PAP).tAdherent stromal cells from day-70 LBMCs were trypsinized,replated on coverslips, and processed for PAP. The percentage ofdonor-originating Glu6PI-a+ cells was <0.01% in the nonadherentcells and in individual CFU-GEMM colonies derived from theseLBMCs.

Proc. Natl. Acad Sci. USA 84 (1987)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 5,

202

1

Proc. Natl. Acad. Sci. USA 84 (1987) 7683

4 e

.9# M

4

FIG. 2. Detection of GB1/6 stromal cell line in adherent cellsexplanted in vitro. (Upper) Adherent marrow cells from the RHL oftransplanted mice at 1 mo after transplantation. (Lower) RHLmarrow cells explanted in vitro from control irradiated but nontrans-planted mice at 1 mo. Donor-originating cells were identified in vitrousing specific rabbit antiserum against murine Glu6PI-a and immu-noperoxidase staining. (x260.) Arrows, Glu6PI-a-containing stromalcells.

mice. The functional integrity of the adherent stromal cellswas quantitated by measuring the longevity of hematopoiesisas cumulative number of total nonadherent cells and multi-

Table 2. Recovery of donor-originating GBlneor cells in culturesexplanted from transplanted mice

G418-resistantStromal colonies per stromal colonies

hind limb*, no. per hind limbt, no.

Group Right Left Right Left

Control-irradiated 65.5 ± 4.1 110.1 ± 10.3 0 0

Transplanted 78.0 ± 13.0 98.7 ± 13.3 30 ± 2.0 10.3 ± 2.0(39 ± 3%) (11 ± 2%)

Mice were transplanted with 5 x 105 GB1neor cells per mouse asdescribed.*Two months after transplantation the total number of cells recov-ered were as follows: 5.8 x 106 per RHL (13 Gy) and 8.6 x 106 perleft hind limb (3 Gy) from control irradiated nontransplanted mice;6.5 x 106 per RHL (13 Gy) and 6.9 x 106 per left hind limb (3 Gy)from transplanted mice. Adherent stromal cell explants wereestablished with 5 x 106 cells per dish (60 x 10 mm). Some were fedbiweekly with G418 (500,ug/ml).Nunmber of G418-resistant colonies were scored 17 days aftercultures were established in vitro. Values in parentheses representpercent control G418-resistant colonies calculated as the number ofG418-resistant stromal cell colonies per 5 x 106 cells by the totalnumber of stromal cell colonies per 5 x 106 cells x 100.

potential progenitor cells produced over 70 days in vitro. Fig.3A shows the cumulative number of viable nonadherent cellsproduced per culture of marrow established at 1, 2, and 3 moafter transplant from the RHL of mice transplanted withGB1/6 cells was higher than that produced by cultures fromirradiated, nontransplanted control mice. The cumulativenumber of multipotential hematopoietic progenitor cellsforming mixed CFU-GEMM colonies per RHL culture (13Gy), established at 1, 2, and 3 mo from transplanted mice was30.5 ± 3.7 x 102, 45.6 ± 2.5 x 102, and 34.7 ± 4.2 x 102,respectively, compared with 5.13 ± 2.2 x 102, 7.3 ± 0.9 x102, and 6.04 ± 0.13 x 102 for control irradiated, nontrans-planted mouse marrow cultures (P < 0.05 for each time-point; Fig. 3B).The effect of donor stromal cell number injected on

recovery of hematopoiesis in irradiated mice was next tested.Fig. 4 shows that LBMCs from the RHL (13 Gy) of irradiatednontransplanted control mice produced 11.45 ± 4.8 x 105cells per flask (Fig. 4A) and 7.41 ± 0.8 x 102 CFU-GEMMprogenitors per flask (Fig. 4C). The cumulative production ofnonadherent cells as well as the CFU-GEMM progenitor cellsin LBMCs from RHL marrow cultures of transplanted miceincreased with the number ofstromal cells injected per mouseover the range of 1 x 105-1 X 106 cells (Fig. 4 A and C).Production of hematopoietic progenitor cells by RHL cul-tures from transplanted mice reached 48% of the level seenin cultures from nonirradiated mice compared with 5% bymarrow cultures from irradiated nontransplanted mice (Fig.

-O 150-0

xu' 1006D

504._CZ

U

~ce._

o

&n

o

-4

L;

A "

s

--10..i

.____

a~.i~~e4 I

40Days in culture

FIG. 3. Hematopoiesis in LBMCs established at 1, 2, and 3 moafter mice were irradiated and transplanted with GB1/6 cell line asdescribed. At monthly intervals at least five mice were sacrificed,and LBMCs were established from the femur and tibia of each righthind limb. Weekly viable nonadherent cells were counted and platedat 5 x 104 cells per ml in the CFU-GEMM assay (28). Results areexpressed as follows: (A) the cumulative mean number of viable cellsproduced per flask over 70 days; and (B) cumulative mean + SEMof mixed and erythroid colony-forming progenitor cells produced inthose same flasks. Irradiated, nonengrafted control mice at 1 mo (i),2 mo (o), and 3 mo (A); GB1/6 transplanted mice at 1 mo (o), 2 mo(o), or 3 mo (A); nonirradiated control mice (x).

.fY_:L

kiiN _i

o s* :tt...-'...i .. _.

k :s ^t o

Medical Sciences: Anklesaria et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 5,

202

1

7684 Medical Sciences: Anklesaria et al.

15C

10(

V)CZ

4-54W)

6r.Vo

151

10

C10

c

._

15

CZ

515

x

i.

o 15

uoc: l

a -

00-L

10io

5

A RHL

B LHL

50 C RHL I1

v0 XS~~~~~~~~~~~

IC'= p .....71

50

50

0

D LHL

II....so-~~~~~~~~~~~~~~~~~~~~

20 40

Days in culture60 80

FIG. 4. Hematopoiesis in LBMCs established from mice trans-planted with different numbers of GB1/6 cells. LBMCs were

established 2 mo after GB1/6 cell transplantation as described forFig. 3. Results are expressed as cumulative mean + SEM of viablecells produced per flask over 70 days, from each RHL (13 Gy) (A) andleft hind limb [LHL (3 Gy); B] cultures and cumulative mean ± SEMof mixed and erythroid CFU-GEMM colony-forming progenitor cellsper flask from RHL (13 Gy) (C) and LHL (3 Gy) (D) cultures. Miceinoculated with 1 x 106 (o), 5 x 105 (A), or 1 x 105 (o) cells; or

irradiated nonengrafted controls (U); control nonirradiated mice (x).

4 A and C). An x-ray dose of 3 Gy to the left hind limbdecreased hematopoietic stem cell production in marrow

cultures from irradiated nontransplanted mice to 35% com-

pared with cultures from nonirradiated mice. However,GB1/6 cell engraftment did not detectably increase cellproduction in LBMCs from limbs irradiated at this dose (Fig.4BandD).The efficiency of repopulation of marrow sinuses of the

RHL (13 Gy) by endogenous CFU-S and the recovery ofperipheral blood counts was next measured in GB1/6 cell-transplanted and in irradiated nontransplanted control mice.Groups ofC57BL/6J mice were irradiated in 2-Gy incrementsfrom 3 Gy to 7 Gy with the RHL receiving between 10 and12.5 Gy (Table 3). A subgroup from each irradiation dosegroup was injected with 5 x 105 GB1/6 cells per mouse(optimal cell inoculum based on data in Fig. 4 A and C). Sixweeks after irradiation and transplantation, the RHL fromeach animal in each dose group was assayed for the numberof multipotential stem cells forming CFU-S. As shown inTable 3 at lower TBI doses of 3 and 5 Gy, the number ofCFU-S-forming multipotential stem cells per RHL was sim-ilar in GB1/6 transplanted and in control irradiated nontrans-planted mice. In contrast, sublethally irradiated (7-Gy TBI)mice transplanted with GB1/6 cells showed a significantlyhigher number of stem cells forming CFU-S in the RHLcompared with the number recovered from the RHL ofirradiated nontransplanted control mice (P < 0.01).The kinetics of recovery of peripheral blood counts in mice

after TBI doses of 3 or 5 Gy were similar in transplanted andcontrol irradiated nontransplanted mice (Table 3). In con-trast, a significant recovery of peripheral blood white bloodcell count (6.9 ± 1.0 x 103 per mm3) and platelet count (112.5± 2.5 x 103 per mm3) was seen in 7-Gy-irradiated GB1/6-transplanted mice compared with irradiated nontransplantedcontrols (white blood cell count: 4 ± 0.05 x 103 cells per mm3;platelet count 50 ± 3 platelets per mm3; P < 0.05; Table 3).

DISCUSSIONThe function of donor-originating stromal cells in the mar-row-transplant recipient is unknown. The data show a stablemurine stromal cell line that provides a favorable microen-vironment for hematopoietic stem cells in vitro, engrafts invivo in irradiated recipients, and stimulates recovery ofrecipient hematopoietic stem cells in vivo. Using specificantiserum against the Glu6PI-a isoenzyme, stromal cells ofdonor origin were identified in marrow sinuses in situ as wellas in adherent cells explanted in vitro several months aftertransplant. Thus, donor-originating stromal cells can infil-

Table 3. Hematopoietic recovery in C57BL/6J mice transplantedwith GB1/6 stromal cell line

Total* Peripheral blood analysisbody CFU-S per on day 42 after irradiationsdose, right hind WBC, PLT, RBC,Gy limb,t no. X 103/mm3 X 103/mm3 X 106/mm33.0 92.0 ± 24.5 8.3 ± 1.5 156.5 ± 6.5 7.7 ± 0.03

(104.6 ± 26.0) (7.2 ± 0.3) (126.5 ± 5.5) (8.2 ± 0.02)5.2 108.3 ± 9.2 6.8 ± 0.3 131.5 ± 10.5 8.2 ± 0.05

(132.5 ± 24.4) (5.7 ± 0.6) (108.0 ± 1.0) (7.4 ± 0.12)7.0 101.8 ± 11.7§ 6.9 ± 1.01 112.5 ± 2.51 7.5 ± 0.5

(49.1 ± 9.5) (4.0 ± 0.05) (50.0 ± 3.0) (5.9 ± 1.3)*Each irradiation group had 5-10 mice per group. All mice receivedan additional 10-12.5 Gy to the RHL. One group was injected with5 x 106 GB1/6 stromal cells per mouse. Values in parentheses arefrom the group that received no cells but were control irradiated.

tSix weeks after irradiation and transplantation, a subgroup fromeach group was sacrificed; cells were flushed from each RHL andassayed for CFU-S (21). Results are expressed as mean ± SD ofthree to five mice. An average of 4.4 ± 2.9 endogenous CFU-Scolonies was seen on the spleens of irradiated noninjected mice.tResults are expressed as mean + SD of at least three mice per group.Nonirradiated mice had a white blood cell (WBC) count of 8.8 ± 1.6x 10/mm3; platelet (PLT) count of 173.7 ± 28.7 x 103/mm3 anderythrocyte (RBC) count of 7.8 + 0.05 x 106/mm3.§P <0.01 compared with values from control irradiated nontrans-planted mice.1p <0.05 in the same comparison as for §.

4

--r

-4

Proc. Natl. Acad. Sci. USA 84 (1987)

T,I

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 5,

202

1

Proc. Natl. Acad. Sci. USA 84 (1987) 7685

trate host marrow sinuses and function in vivo to supportrecovery of primitive hematopoietic stem cells includingthose forming CFU-S and CFU-GEMM. Whether stimula-tion of CFU-S recovery by GB1/6 cells in inoculated micecan be attributed to stimulation ofCFUs at the irradiation sitein proximity to seeded GB1/6 cells, to homing of CFU-Sarriving via the circulation from other sites, or to bothmechanisms is unknown.The GB1/6 cell line was chosen for these studies because

of its endothelial-like characteristics and in vitro supportcapacity for multipotential hematopoietic progenitor cells.Constitutive production of M-CSF and no detectable pro-duction of IL-3, GM-CSF, nor G-CSF suggests that itssupport capacity for stem cells may be due either to the abilityof M-CSF to trigger release of other CSFs from accessorycells in vivo and in vitro (24) or to another growth factor.Other clonal stromal cell lines demonstrated by us (11) andothers (25) do not support multipotential hematopoieticprogenitor cells forming CFU-S in vitro. Whether these othercell lines also engraft in vivo and support or suppresshematopoietic recovery after irradiation is unclear.Marrow cultures established in vitro from the heavily

irradiated RHL (13 Gy) ofmice transplanted with GB1/6 cellsshowed stable engraftment of donor stromal cells over 3 moand enhanced hematopoietic progenitor cell production invitro in LBMCs compared with cultures derived from irra-diated nontransplanted control mice. In contrast, hemato-poiesis in LBMCs established from the 3-Gy-irradiated lefthind limb of transplanted mice was not enhanced comparedwith cultures from irradiated nontransplanted control mice,and donor-originating stromal cells identified at 1 mo did notpersist. The data confirm that previous irradiation damagesthe stromal cells of the hematopoietic microenvironment (26)and support Brecher et al. (27), who suggested that niches,freed of endogenous hematopoietic stem cells by the higherdose, provide more efficient seeding sites for injected donorcells. Our data may explain previous failure to detect donor-originating stromal cells in studies using lower x-ray doses (6,8).To establish a detectable chimeric stromal cell population,

a minimum of 1 x 105 purified GB1/6 stromal cells wererequired. Thus, other studies reporting no donor-originatingstromal cells in LBMC adherent cell layers derived fromtransplanted mice using chromosomally marked marrow cellsmay not have used enough transplanted cells.The physiological failure of bone marrow has generally

been attributed to defects in the hematopoietic stem cells-defects correctable by autologous or allogenic bone marrow-stem cell transplantation (28). However, the pathophysiologyof some forms of marrow failure, as in chronic myelogenousleukemia (29) and aplastic anemia (30, 31), may be due to adefect in the stromal microenvironment. Our data revealedevidence for stimulation of hematopoietic recovery in vivo instromal cell line-engrafted mice compared with irradiatednontransplanted controls. Although further studies are re-quired, a therapeutic role for stromal cell infusion in con-junction with hematopoietic stem cell transplantation issuggested in the treatment of some forms of marrow failure.

We thank Cetus Corporation for supplying plasmid-containingmurine CSF-1 cDNA, Dr. A. Webb for plasmid-containing IL-1f3cDNA, and Dr. S. Nagata for plasmid-containing murine G-CSFcDNA. Cell lines MBA-1, MBA-13, and 14F-1 were generously

supplied by Dr. Dov Zipori of Weizmann Institute, Israel. Antiserumto collagen types I and IV were a gift from Dr. Hynda Kleinman. Thiswork was supported by Grants P01HD19767 and R01CA39851 fromthe National Institutes of Health to J.S.G.

1. Chamberlin, W., Barone, J., Kedo, A. & Fried, W. (1974) Blood 44,385-392.

2. Werts, E. D., Gibson, D. P., Knapp, S. A. & DeGowin, R. L.(1980) Radiat. Res. 81, 20-30.

3. Piersma, A. H., Ploemacher, R. E. & Brockbank, K. G. M. (1983)Br. J. Haematol. 54, 285-290.

4. Marshall, M. J., Nisket, N. W. & Eaven, S. (1984) Experientia 40,385-386.

5. Keating, A., Singer, J. W., Killen, P. D., Striker, G. E., Salo,A. C., Sanders, J., Thomas, E. D., Thorning, D. & Fialkow, P. J.(1982) Nature (London) 298, 280-283.

6. Chertkov, J. L., Drize, N. J., Gurevitch, 0. A. & Samoylova,R. S. (1985) Exp. Hematol. (NY) 13, 1217-1222.

7. Golde, D. W., Hocking, W. G., Quan, S. G., Sparkes, R. S. &Gale, R. P. (1980) Br. J. Haematol. 44, 183-187.

8. Lennon, J. E. & Micklem, H. S. (1986) Exp. Hematol. (NY) 14,287-292.

9. Greenberger, J. S. (1978) Nature (London) 275, 752-754.10. Anklesaria, P. N., Sakakeeny, M. A., Klassen, V., Rothstein, L.,

FitzGerald, T. J., Greenberger, J. S. & Holland, C. (1987) Exp.Hematol. (NY) 15, 195-202.

11. Anklesaria, P. N., Klassen, V., Sakakeeny, M. A., FitzGerald,T. J., Harrison, D., Rybak, M. E. & Greenberger, J. S. (1987) Exp.Hematol. (NY) 15, 636-644.

12. Pearse, A. G. E. (1986) Histochemistry (Churchill, London), Vol.1, 3rd Ed.

13. Charles, D. & Lee, C.-Y. (1980) Mol. Cell. Biochem. 29, 11-21.14. Hadji, M. & Morales, A. R. (1983) Lab Med. 14, 767-771.15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory,Cold Spring Harbor, NY).

16. Fung, M. C., Hapel, A. J., Ymer, S., Cohen, D. R., Johnson,R. M., Campbell, M. D. & Young, I. G. (1984) Nature (London)307, 233-237.

17. Gough, N., Gough, J., Metcalf, D., Kelow, A., Groel, D., Nicola,N., Burgess, A. & Dunn, A. (1984) Nature (London) 309, 763-767.

18. Kawasaki, E. S., Lander, M. B., Wang, A. M., VanArsdell, J.,Warren, K., Coyne, M. Y., Schweickart, V. L., Lee, M. T.,Wilson, K. J., Boosman, A., Stanley, R., Ralph, P. & Mark, D. F.(1985) Science 230, 291-2%.

19. Auron, P. E., Webb, A. C., Rosenwasser, L. J., Mucci, S. F.,Rich, A., Wolff, S. M. & Dinarello, C. A. (1984) Proc. Natl. Acad.Sci. USA 81, 7907-7911.

20. Isuchiya, M., Asano, S., Kaziro, Y. & Nagata, S. (1986) Proc.Natl. Acad. Sci. USA 83, 7633-7637.

21. Magli, M. C., Iscove, N. N. & Odartchenko, N. (1982) Nature(London) 295, 527-529.

22. Humphries, R. K., Eaves, A. C. & Eaves, C. J. (1981) Proc. Natl.Acad. Sci. USA 78, 3629-3633.

23. Zipori, D., Duskin, D., Tamir, M., Argaman, A., Toledo, J. K. &Malik, Z. (1985) J. Cell. Physiol. 122, 81-90.

24. Broxmeyer, H. E., Williams, D. E., Cooper, S., Waheed, A. &Shadduck, R. K. (1987) Blood 69, 913-918.

25. Rennick, D., Yang, G., Gemmell, L. & Lee, F. (1987) Blood 69,682-691.

26. Greenberger, J. S., Eckner, R. J., Otten, J. A. & Tennant, R. W.(1982) Int. J. Radiat. Oncol. Biol. Phys. 8, 1155-1165.

27. Brecher, G., Tjio, J. H., Haley, J. E., Narla, J. & Beal, S. L.(1979) Blood Cells 5, 237-246.

28. Ramsay, N., LeBien, T., Nesbit, M., McGlave, P., Weisdorf, D.,Kenyon, P., Hurd, D., Goldman, A., Kim, T. & Kersey, J. (1985)Blood 66, 508-513.

29. Takahashi, M., Keating, A. & Singer, J. W. (1985) Exp. Hematol.(NY) 13, 926-931.

30. Juneja, H. S., Gardner, F. M., Minguell, J. J. & Helmer, R. E.(1984) Exp. Hematol. (NY) 12, 221-230.

31. Champlin, R. E., Feig, S. A. & Gale, R. P. (1984) Exp. Hematol.(NY) 12, 728-733.

Medical Sciences: Anklesaria et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 5,

202

1